EDA notebook

Loading Packages

library(readr)
library(tibble)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(tidyverse)

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ forcats   1.0.0     ✔ purrr     1.0.2
## ✔ ggplot2   3.5.1     ✔ stringr   1.5.1
## ✔ lubridate 1.9.3     ✔ tidyr     1.3.1

## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

library(ggplot2)
data_replic <- read_csv("/Users/soniadalal/Desktop/302W/302W Project/data/data and replication code/data_replic.csv")

## New names:
## Rows: 1012 Columns: 27
## ── Column specification
## ──────────────────────────────────────────────────────── Delimiter: "," chr
## (3): iso3, region, income_group dbl (22): year, homrates_unodc, homrates_who,
## deport_convict, deport_convict... lgl (2): ...19, ...23
## ℹ Use `spec()` to retrieve the full column specification for this data. ℹ
## Specify the column types or set `show_col_types = FALSE` to quiet this message.
## • `` -> `...19`
## • `` -> `...23`

head(data_replic)

## # A tibble: 6 × 27
##   iso3   year region     income_group homrates_unodc homrates_who deport_convict
##   <chr> <dbl> <chr>      <chr>                 <dbl>        <dbl>          <dbl>
## 1 AFG    2010 South Asia Low income             3.81        NA            0.0194
## 2 AFG    2012 South Asia Low income             5.41        NA            0.0265
## 3 ALB    1996 Europe & … Upper middl…          NA            8.36         0.142 
## 4 ALB    1998 Europe & … Upper middl…          NA           31.7          0.159 
## 5 ALB    2000 Europe & … Upper middl…          NA           10.7          0.339 
## 6 ALB    2002 Europe & … Upper middl…          NA            7.13         0.393 
## # ℹ 20 more variables: deport_convict_d <dbl>, deport_convict_lead <dbl>,
## #   deport_nonconvict <dbl>, deport_nonconvict_d <dbl>, corruption <dbl>,
## #   crime_US_weight <dbl>, GDP_growth <dbl>,
## #   GDPpercapita_const2010USD_log <dbl>, gini_mkt <dbl>,
## #   instr_benef_medicaidpregn_lag1 <dbl>, instr_enforce_everify_lag1 <dbl>,
## #   ...19 <lgl>, polity2 <dbl>, pop_sh14 <dbl>, population_log <dbl>,
## #   ...23 <lgl>, remittances_GDP <dbl>, urban_interp <dbl>, …

summary(data_replic)

##      iso3                year         region          income_group      
##  Length:1012        Min.   :1996   Length:1012        Length:1012       
##  Class :character   1st Qu.:2002   Class :character   Class :character  
##  Mode  :character   Median :2006   Mode  :character   Mode  :character  
##                     Mean   :2006                                        
##                     3rd Qu.:2010                                        
##                     Max.   :2014                                        
##                                                                         
##  homrates_unodc    homrates_who    deport_convict      deport_convict_d    
##  Min.   : 0.280   Min.   : 0.000   Min.   :  0.00000   Min.   :-164.00466  
##  1st Qu.: 1.394   1st Qu.: 1.092   1st Qu.:  0.04974   1st Qu.:  -0.04479  
##  Median : 3.390   Median : 2.805   Median :  0.17095   Median :   0.00127  
##  Mean   : 8.731   Mean   : 7.190   Mean   :  6.02820   Mean   :   0.34473  
##  3rd Qu.: 9.279   3rd Qu.: 9.119   3rd Qu.:  1.14709   3rd Qu.:   0.07173  
##  Max.   :92.960   Max.   :74.846   Max.   :175.44412   Max.   :  57.59695  
##  NA's   :398      NA's   :208      NA's   :9           NA's   :53          
##  deport_convict_lead  deport_nonconvict deport_nonconvict_d   corruption   
##  Min.   :-164.00466   Min.   :  0.000   Min.   :-117.9508   Min.   :0.000  
##  1st Qu.:  -0.05009   1st Qu.:  0.120   1st Qu.:  -0.1497   1st Qu.:2.000  
##  Median :   0.00000   Median :  0.392   Median :   0.0000   Median :2.500  
##  Mean   :   0.19557   Mean   :  7.350   Mean   :   0.9367   Mean   :2.981  
##  3rd Qu.:   0.07078   3rd Qu.:  2.470   3rd Qu.:   0.1312   3rd Qu.:4.000  
##  Max.   :  57.59695   Max.   :307.606   Max.   : 196.3743   Max.   :6.000  
##  NA's   :132          NA's   :143       NA's   :155         NA's   :204    
##  crime_US_weight   GDP_growth      GDPpercapita_const2010USD_log
##  Min.   :341.5   Min.   :-16.700   Min.   : 5.755               
##  1st Qu.:407.8   1st Qu.:  1.730   1st Qu.: 8.082               
##  Median :466.6   Median :  3.852   Median : 8.915               
##  Mean   :464.1   Mean   :  3.900   Mean   : 8.949               
##  3rd Qu.:506.0   3rd Qu.:  5.970   3rd Qu.:10.056               
##  Max.   :724.9   Max.   : 34.500   Max.   :11.485               
##  NA's   :234     NA's   :12        NA's   :16                   
##     gini_mkt     instr_benef_medicaidpregn_lag1 instr_enforce_everify_lag1
##  Min.   :27.90   Min.   :0.1484                 Min.   :0.00000           
##  1st Qu.:43.00   1st Qu.:0.6028                 1st Qu.:0.00000           
##  Median :46.90   Median :0.7047                 Median :0.03464           
##  Mean   :46.53   Mean   :0.6826                 Mean   :0.07168           
##  3rd Qu.:50.10   3rd Qu.:0.7784                 3rd Qu.:0.09822           
##  Max.   :68.50   Max.   :0.9633                 Max.   :0.79440           
##  NA's   :132     NA's   :313                    NA's   :313               
##   ...19            polity2          pop_sh14     population_log  
##  Mode:logical   Min.   :-9.000   Min.   :11.06   Min.   : 9.678  
##  NA's:1012      1st Qu.: 5.000   1st Qu.:17.57   1st Qu.:15.036  
##                 Median : 8.000   Median :24.05   Median :15.958  
##                 Mean   : 5.822   Mean   :25.39   Mean   :15.819  
##                 3rd Qu.:10.000   3rd Qu.:31.72   3rd Qu.:17.229  
##                 Max.   :10.000   Max.   :49.87   Max.   :21.034  
##                 NA's   :153      NA's   :37                      
##   ...23         remittances_GDP     urban_interp    war_intrastate    
##  Mode:logical   Min.   : 0.00139   Min.   : 10.30   Min.   :0.000000  
##  NA's:1012      1st Qu.: 0.29699   1st Qu.: 49.17   1st Qu.:0.000000  
##                 Median : 1.39941   Median : 64.65   Median :0.000000  
##                 Mean   : 3.95100   Mean   : 62.13   Mean   :0.006917  
##                 3rd Qu.: 4.83025   3rd Qu.: 77.05   3rd Qu.:0.000000  
##                 Max.   :49.28990   Max.   :100.00   Max.   :1.000000  
##                 NA's   :86                                            
##      yr_sch      
##  Min.   : 2.870  
##  1st Qu.: 7.432  
##  Median : 9.399  
##  Mean   : 9.029  
##  3rd Qu.:10.810  
##  Max.   :16.815  
##  NA's   :183

This dataset shows a comprehensive set of variables with demographic, economic, and crime-related metrics across different countries over time. The UNODC recorded homicide rates and deportations of convicts can be used to determine the impact of deportations on crime rates in different regions of home countries. This relationship can also be analyzed through a socioeconomic lense using the income data provided

iso3: ISO 3166-1 alpha-3 country code representing each country. year: Year of the data record. region: World Bank-defined global region that the country belongs to. income_group: World Bank-defined income category of the country (e.g., Low income, Upper middle income). homrates_unodc: Homicide rates per 100,000 population as reported by the United Nations Office on Drugs and Crime. homrates_who: Homicide rates per 100,000 population as reported by the World Health Organization. deport_convict: Deportation rates of convicted individuals per 100,000 population. deport_convict_d: Differenced series of deportation rates of convicted individuals, indicating change from the previous year. deport_convict_lead: Lead series of deportation rates of convicted individuals, indicating the value for the forthcoming year. deport_nonconvict: Deportation rates of non-convicted individuals per 100,000 population. corruption: Higher corruption index indicates lower rates of corruption instr_enforce_everify_lag1: Previous year’s enforcement measure related to electronic verification systems. polity2: Score from the Polity IV project measuring the level of democracy and autocracy. pop_sh14: Percentage of the population aged under 14 years. population_log: Logarithm of the population size to normalize the data. remittances_GDP: Remittances received as a percentage of the GDP. urban_interp: Urbanization rate interpolated for non-census years. war_intrastate: Indicator of whether there was an intrastate war occurring. yr_sch: Average years of schooling in the population.

Remove rows with missing values in key columns, remove NA columns.

df <- data_replic %>%
  select(-...19, -...23)

newdf <- data_replic %>%
  select(-...19, -...23) %>%
  filter(!is.na(deport_nonconvict) & !is.na(homrates_unodc) & !is.na(region) & !is.na(deport_convict) & !is.na(income_group)) 
newdf

## # A tibble: 537 × 25
##    iso3   year region    income_group homrates_unodc homrates_who deport_convict
##    <chr> <dbl> <chr>     <chr>                 <dbl>        <dbl>          <dbl>
##  1 AFG    2010 South As… Low income             3.81        NA            0.0194
##  2 ALB    2006 Europe &… Upper middl…           4.06         0            1.07  
##  3 ALB    2008 Europe &… Upper middl…           3.31         2.87         0.964 
##  4 ALB    2010 Europe &… Upper middl…           3.64         1.15         1.59  
##  5 ALB    2012 Europe &… Upper middl…           5.18        NA            1.27  
##  6 ALB    2014 Europe &… Upper middl…           4.18        NA            0.916 
##  7 ARM    2006 Europe &… Lower middl…           2.28        NA            0.993 
##  8 ARM    2008 Europe &… Lower middl…           2.92        NA            1.32  
##  9 ARM    2010 Europe &… Lower middl…           2.58         1.70         1.04  
## 10 ARM    2012 Europe &… Lower middl…           2.28         1.61         1.18  
## # ℹ 527 more rows
## # ℹ 18 more variables: deport_convict_d <dbl>, deport_convict_lead <dbl>,
## #   deport_nonconvict <dbl>, deport_nonconvict_d <dbl>, corruption <dbl>,
## #   crime_US_weight <dbl>, GDP_growth <dbl>,
## #   GDPpercapita_const2010USD_log <dbl>, gini_mkt <dbl>,
## #   instr_benef_medicaidpregn_lag1 <dbl>, instr_enforce_everify_lag1 <dbl>,
## #   polity2 <dbl>, pop_sh14 <dbl>, population_log <dbl>, …

#Make a correlation plot

install.packages("corrplot")

## The following package(s) will be installed:
## - corrplot [0.92]
## These packages will be installed into "~/Desktop/302W/302W Project/data/data and replication code/renv/library/R-4.2/x86_64-apple-darwin17.0".
## 
## # Installing packages --------------------------------------------------------
## - Installing corrplot ...                       OK [linked from cache]
## Successfully installed 1 package in 11 milliseconds.

library(corrplot)

## corrplot 0.92 loaded

df <- df %>%
  mutate(
    income_group_numeric = case_when(
      income_group == "Low income" ~ 1,
      income_group == "Lower middle income" ~ 2,
      income_group == "Upper middle income" ~ 3,
      income_group == "High income: nonOECD" ~ 4,
      income_group == "High income: OECD" ~ 5,
      TRUE ~ NA_real_  # Handles any cases that don't match the above
    )
  )

General overview of correlations between different variables

# Select relevant variables and remove NA values for correlation analysis
cor_data <- df %>%
  select(corruption, homrates_unodc, GDPpercapita_const2010USD_log, urban_interp, year, deport_convict, deport_nonconvict, polity2, yr_sch, income_group_numeric) %>%
  na.omit()  # Remove rows with NA values

# Calculate the correlation matrix
cor_matrix <- cor(cor_data)
corrplot(cor_matrix, method = "color", type = "upper", order = "hclust",
         addCoef.col = "black",  # Add correlation coefficients to enhance readability
         tl.col = "black", tl.srt = 45,  # Adjust text label color and rotation for better visibility
         diag = FALSE,  # Exclude diagonal elements to avoid redundancy
         number.cex = 0.8,  # Adjust size of correlation coefficients
         tl.cex = 0.8)

Shows that as deportation increases, homicide rates increase but over the years available in the data, homicide rates and deportations rates did not increase significantly.

#Income distribution Counts in dataset for each region

# Bar plot
ggplot(df, aes(x = region, fill = income_group)) +
  geom_bar(position = "dodge") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(title = "Income Distribution Counts for Each Region")

More data on income group is available for Latin America & Caribbean and Europe and Central Asia than other regions. #Convict Deportation distribution in dataset by Region

# Bar plot
ggplot(df, aes(x = region, fill = deport_convict)) +
  geom_bar(position = "dodge") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(title = "Convict Deportation Rate Distribution Counts per Region")

## Warning: The following aesthetics were dropped during statistical transformation: fill.
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
##   the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
##   variable into a factor?

More data on convict deportation is available for Latin America & Caribbean and Europe and Central Asia than other regions. #Non-Convict Deportation distribution in dataset by Region

# Bar plot
ggplot(df, aes(x = region, fill = deport_nonconvict)) +
  geom_bar(position = "dodge") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(title = "Non-Convict Deportation Rate Distribution Counts per Region")

## Warning: The following aesthetics were dropped during statistical transformation: fill.
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
##   the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
##   variable into a factor?

More data on deportation of nonconvicts is available for Latin America & Caribbean and Europe and Central Asia than other regions. #Corruption distribution in dataset by Region

# Bar plot
ggplot(df, aes(x = region, fill = corruption)) +
  geom_bar(position = "dodge") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(title = "Corruption Score Distribution Counts per Region")

## Warning: The following aesthetics were dropped during statistical transformation: fill.
## ℹ This can happen when ggplot fails to infer the correct grouping structure in
##   the data.
## ℹ Did you forget to specify a `group` aesthetic or to convert a numerical
##   variable into a factor?

More data on corruption is available for Latin America & Caribbean and Europe and Central Asia than other regions. # Boxplot of Corruption Scores for all the Regions

# Filter the dataset to include only relevant variables

corruption_data <- data_replic %>%
  select(region, corruption)

# Summary statistics by region
summary_stats <- corruption_data %>%
  group_by(region) %>%
  summarise(mean_corruption = mean(corruption, na.rm = TRUE),
            median_corruption = median(corruption, na.rm = TRUE),
            min_corruption = min(corruption, na.rm = TRUE),
            max_corruption = max(corruption, na.rm = TRUE))

# Visualization: Box plot of corruption scores by region
boxplot_corruption <- ggplot(corruption_data, aes(x = region, y = corruption, fill = region)) +
  geom_boxplot() +
  labs(title = "Corruption Scores by Region",
       x = "Region",
       y = "Corruption Score") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))  # Rotate x-axis labels for better readability

# Display the summary statistics and box plot
print(summary_stats)

## # A tibble: 8 × 5
##   region         mean_corruption median_corruption min_corruption max_corruption
##   <chr>                    <dbl>             <dbl>          <dbl>          <dbl>
## 1 East Asia & P…            3.24               3              1              5.5
## 2 Europe & Cent…            3.5                3              1              6  
## 3 Latin America…            2.49               2.5            1              5  
## 4 Middle East &…            2.39               2              1              5  
## 5 North America             4.88               5              3.5            6  
## 6 South Asia                2.48               2.5            1              4  
## 7 Sub-Saharan A…            1.93               2              0              3  
## 8 <NA>                      2.5                2.5            2              3

print(boxplot_corruption)

## Warning: Removed 204 rows containing non-finite outside the scale range
## (`stat_boxplot()`).

Latin America & Caribbean have more corruption on average than Europe & Central Asia #Income distribution proportions for each region in data with no NA for deportation or homicides.

# Calculate the count per income group within each region
newdf_with_count <- newdf %>%
  group_by(region, income_group) %>%
  summarise(count = n(), .groups = 'drop')
new

## function (Class, ...) 
## {
##     ClassDef <- getClass(Class, where = topenv(parent.frame()))
##     value <- .Call(C_new_object, ClassDef)
##     initialize(value, ...)
## }
## <bytecode: 0x7f7ffc275170>
## <environment: namespace:methods>

# Calculate the total count per region
totals <- newdf_with_count %>%
  group_by(region) %>%
  summarise(total = sum(count), .groups = 'drop')

# Join the totals back to the original data frame
newdf_with_proportions <- newdf_with_count %>%
  left_join(totals, by = "region") %>%
  mutate(proportion = count / total)

# Plot the proportions instead of count
ggplot(newdf_with_proportions, aes(x = region, y = proportion, fill = income_group)) +
  geom_bar(position = "dodge", stat = "identity") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  ylab("Proportion")

Europe & Central Asia has a higher proportion of High income whereas Latin America & Caribbean has a higher proportion of Upper middle income.

#Income distribution proportions for each region with darker colors corresponding to lower income

# Order the income_group factor based on income levels from highest to lowest
newdf_with_proportions$income_group <- factor(
  newdf_with_proportions$income_group, 
  levels = c(
    "Low income",
    "Lower middle income",
    "Upper middle income",
    "High income: nonOECD",
    "High income: OECD" 
    
  ),
  ordered = TRUE
)
ggplot(newdf_with_proportions, aes(x = region, y = proportion, fill = income_group)) +
  geom_bar(stat = "identity", position = position_dodge(preserve = 'single')) +
  scale_fill_brewer(palette = "Blues", direction = -1) + # Use reversed Blues palette for darker colors on lower incomes
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  labs(y = "Proportion", fill = "Income Group")

# Correlational plot with Convict Deportation Rates and Homicide Rates by Region (Scaled and Unscaled)

ggplot(newdf, aes(x = deport_convict, y = homrates_unodc)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region, scale = "free") +
  labs(title = "Correlation Between Convict Deportation Rates and Homicide Rates by Region",
       x = "Convict Deportation Rate",
       y = "UNODC Homicide Rate")

## `geom_smooth()` using formula = 'y ~ x'

ggplot(newdf, aes(x = deport_convict, y = homrates_unodc)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region) +
  labs(title = "Correlation Between Convict Deportation Rates and Homicide Rates by Region",
       x = "Non-Convict Deportation Rate",
       y = "UNODC Homicide Rate")

## `geom_smooth()` using formula = 'y ~ x'

correlation_by_region <- newdf %>%
  group_by(region) %>%
  summarize(correlation = cor(deport_convict, homrates_unodc, use = "complete.obs"))

print(correlation_by_region)

## # A tibble: 7 × 2
##   region                     correlation
##   <chr>                            <dbl>
## 1 East Asia & Pacific             0.101 
## 2 Europe & Central Asia           0.0529
## 3 Latin America & Caribbean       0.616 
## 4 Middle East & North Africa     -0.0952
## 5 North America                   0.996 
## 6 South Asia                      0.493 
## 7 Sub-Saharan Africa             -0.0725

Correlational plot with Non-Convict Deportation Rates and Homicide Rates by Region (Scaled and Unscaled)

ggplot(newdf, aes(x = deport_nonconvict, y = homrates_unodc)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region, scale = "free") +
  labs(title = "Correlation Between Non-Convict Deportation Rates and Homicide Rates by Region",
       x = "Non-Convict Deportation Rate",
       y = "UNODC Homicide Rate")

## `geom_smooth()` using formula = 'y ~ x'

ggplot(newdf, aes(x = deport_nonconvict, y = homrates_unodc)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region) +
  labs(title = "Correlation Between Non-Convict Deportation Rates and Homicide Rates by Region",
       x = "Non-Convict Deportation Rate",
       y = "UNODC Homicide Rate")

## `geom_smooth()` using formula = 'y ~ x'

correlation_by_region2 <- newdf %>%
  group_by(region) %>%
  summarize(correlation = cor(deport_nonconvict, homrates_unodc, use = "complete.obs"))

print(correlation_by_region2)

## # A tibble: 7 × 2
##   region                     correlation
##   <chr>                            <dbl>
## 1 East Asia & Pacific              0.158
## 2 Europe & Central Asia            0.132
## 3 Latin America & Caribbean        0.576
## 4 Middle East & North Africa      -0.142
## 5 North America                    0.890
## 6 South Asia                       0.364
## 7 Sub-Saharan Africa              -0.178

These correlational plots show the most comprehensive data for non-convicts and convicts for Latin American and the Carribean. For this reason we will forcus on this region. # Distribution of Years in the Dataset

# Preparing the data by ensuring 'year' is available and correctly formatted
newdf <- mutate(newdf, year = as.integer(year))

# Combine into one dataframe for plotting

# Plotting the histogram of year distributions across datasets
ggplot(newdf, aes(x = year)) +
  geom_histogram(position = "identity", alpha = 0.5, binwidth = 1) +
  labs(title = "Distribution of Years", x = "Year", y = "Count") +
  scale_fill_manual(values = c("blue", "red", "green")) +
  theme_minimal()

Data only available from 2004 to 2014. This plus the correlation plot shows a limitation that there hasn’t been much change in the variables which may make it more difficult to find relationships in this time.

# Filter the dataset for Latin America & Caribbean region
latin_america_data <- newdf %>%
  filter(region == "Latin America & Caribbean")

# Impute missing data for odd years using linear interpolation
imputed_data1 <- latin_america_data %>%
  tidyr::complete(year = seq(min(year), max(year), by = 2)) %>%
  mutate(homrates_unodc = approx(year[!is.na(homrates_unodc)], homrates_unodc[!is.na(homrates_unodc)], year)$y)

## Warning: There was 1 warning in `mutate()`.
## ℹ In argument: `homrates_unodc = `$`(...)`.
## Caused by warning in `regularize.values()`:
## ! collapsing to unique 'x' values

# Create the line plot
ggplot(imputed_data1, aes(x = year, y = homrates_unodc)) +
  geom_line() +
  labs(title = "Homicide Rates in Latin America & Caribbean",
       x = "Year",
       y = "Homicide Rates (per 100,000 population)") +
  theme_minimal()

# Filter the dataset for Latin America & Caribbean region
europe_central_asia_data <- newdf %>%
  filter(region == "Europe & Central Asia")

# Impute missing data for odd years using linear interpolation
imputed_data2 <- europe_central_asia_data %>%
  tidyr::complete(year = seq(min(year), max(year), by = 2)) %>%
  mutate(homrates_unodc = approx(year[!is.na(homrates_unodc)], homrates_unodc[!is.na(homrates_unodc)], year)$y)

## Warning: There was 1 warning in `mutate()`.
## ℹ In argument: `homrates_unodc = `$`(...)`.
## Caused by warning in `regularize.values()`:
## ! collapsing to unique 'x' values

# Create the line plot
ggplot(imputed_data2, aes(x = year, y = homrates_unodc)) +
  geom_line() +
  labs(title = "Homicide Rates in Europe and Central Asia",
       x = "Year",
       y = "Homicide Rates (per 100,000 population)") +
  theme_minimal()

combined_data <- rbind(imputed_data1, imputed_data2)
ggplot(combined_data, aes(x = year, y = homrates_unodc, color = region)) +
  geom_line() +
  labs(title = "Homicide Rates in Latin America & Caribbean vs. Europe & Central Asia",
       x = "Year",
       y = "Homicide Rates (per 100,000 population)") +
  theme_minimal() +
  scale_color_manual(values = c("blue", "red"))

Europe and Central Asia Homicide rates are low and slightly decrease over the years while homicide rates in Latin America and the Caribbean are much higher and increase then decrease. # Make new dataset with just Latin America and the Caribbean

# Filter the dataset for Latin America & Caribbean region
latin_america_data <- newdf %>%
  filter(region == "Latin America & Caribbean")

# Impute missing data for odd years using linear interpolation
imputed_data <- latin_america_data %>%
  tidyr::complete(year = seq(min(year), max(year), by = 2)) %>%
  mutate(homrates_unodc = approx(year[!is.na(homrates_unodc)], homrates_unodc[!is.na(homrates_unodc)], year)$y,
         deport_convict = approx(year[!is.na(deport_convict)], deport_convict[!is.na(deport_convict)], year)$y,
         deport_nonconvict = approx(year[!is.na(deport_nonconvict)], deport_nonconvict[!is.na(deport_nonconvict)], year)$y)

## Warning: There were 3 warnings in `mutate()`.
## The first warning was:
## ℹ In argument: `homrates_unodc = `$`(...)`.
## Caused by warning in `regularize.values()`:
## ! collapsing to unique 'x' values
## ℹ Run `dplyr::last_dplyr_warnings()` to see the 2 remaining warnings.

# Create the line plot
ggplot(imputed_data, aes(x = year, y = homrates_unodc)) +
  geom_line(aes(color = "Homicide Rates"), size = 1) +
  geom_line(aes(y = deport_convict, color = "Convict Deportation Rates"), linetype = "dashed", size = 1) +
  geom_line(aes(y = deport_nonconvict, color = "Non-Convict Deportation Rates"), linetype = "dotdash", size = 1) +
  labs(title = "Homicide and Deportation Rates in Latin America & Caribbean",
       x = "Year",
       y = "Rates per 100,000 population",
       color = "Legend") +
  scale_color_manual(values = c("Homicide Rates" = "red",
                                "Convict Deportation Rates" = "blue",
                                "Non-Convict Deportation Rates" = "green")) +
  theme_minimal() +
  theme(legend.title = element_blank())

## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

# Filter data for Europe & Central Asia
europe_data <- newdf %>%
  filter(region == "Europe & Central Asia")

# Impute missing data for odd years using linear interpolation
imputed_europe_data <- europe_data %>%
  tidyr::complete(year = seq(min(year), max(year), by = 2)) %>%
  mutate(homrates_unodc = approx(year[!is.na(homrates_unodc)], homrates_unodc[!is.na(homrates_unodc)], year)$y,
         deport_convict = approx(year[!is.na(deport_convict)], deport_convict[!is.na(deport_convict)], year)$y,
         deport_nonconvict = approx(year[!is.na(deport_nonconvict)], deport_nonconvict[!is.na(deport_nonconvict)], year)$y)

## Warning: There were 3 warnings in `mutate()`.
## The first warning was:
## ℹ In argument: `homrates_unodc = `$`(...)`.
## Caused by warning in `regularize.values()`:
## ! collapsing to unique 'x' values
## ℹ Run `dplyr::last_dplyr_warnings()` to see the 2 remaining warnings.

# Create the plot for Europe & Central Asia
plot_europe <- ggplot(imputed_europe_data) +
  geom_line(aes(x = year, y = homrates_unodc, color = "Homicide Rates"), size = 1) +
  geom_line(aes(x = year, y = deport_convict, color = "Convict Deportation Rates"), size = 1, linetype = "dashed") +
  geom_line(aes(x = year, y = deport_nonconvict, color = "Non-Convict Deportation Rates"), size = 1, linetype = "dotdash") +
  labs(title = "Homicide and Deportation Rates in Europe & Central Asia",
       x = "Year",
       y = "Rates per 100,000 population",
       color = "Legend") +
  scale_color_manual(values = c("Homicide Rates" = "red",
                                "Convict Deportation Rates" = "blue",
                                "Non-Convict Deportation Rates" = "green")) +
  theme_minimal() +
  theme(legend.title = element_blank())

# Display the plot for Europe & Central Asia
plot_europe

latin_america_tot <- df %>%
  filter(region == "Latin America & Caribbean")

imputed_latin_america_tot <- latin_america_tot %>%
  tidyr::complete(year = seq(min(year), max(year), by = 2)) %>%
  mutate(deport_convict = approx(year[!is.na(deport_convict)], deport_convict[!is.na(deport_convict)], year)$y,
         deport_nonconvict = approx(year[!is.na(deport_nonconvict)], deport_nonconvict[!is.na(deport_nonconvict)], year)$y)

## Warning: There were 2 warnings in `mutate()`.
## The first warning was:
## ℹ In argument: `deport_convict = `$`(...)`.
## Caused by warning in `regularize.values()`:
## ! collapsing to unique 'x' values
## ℹ Run `dplyr::last_dplyr_warnings()` to see the 1 remaining warning.

ggplot(imputed_latin_america_tot) +
  geom_line(aes(x = year, y = deport_convict, color = "Convict Deportation Rates"), size = 1) +
  geom_line(aes(x = year, y = deport_nonconvict, color = "Non-Convict Deportation Rates"), size = 1) +
  labs(title = "Deportation Rates in Latin America & Caribbean",
       x = "Year",
       y = "Rates per 100,000 population",
       color = "Legend") +
  scale_color_manual(values = c("Convict Deportation Rates" = "blue",
                                "Non-Convict Deportation Rates" = "green")) +
  theme_minimal() +
  theme(legend.title = element_blank())

europe_tot <- df %>%
  filter(region == "Europe & Central Asia")

imputed_europe_tot <- europe_tot %>%
  tidyr::complete(year = seq(min(year), max(year), by = 2)) %>%
  mutate(deport_convict = approx(year[!is.na(deport_convict)], deport_convict[!is.na(deport_convict)], year)$y,
         deport_nonconvict = approx(year[!is.na(deport_nonconvict)], deport_nonconvict[!is.na(deport_nonconvict)], year)$y)

## Warning: There were 2 warnings in `mutate()`.
## The first warning was:
## ℹ In argument: `deport_convict = `$`(...)`.
## Caused by warning in `regularize.values()`:
## ! collapsing to unique 'x' values
## ℹ Run `dplyr::last_dplyr_warnings()` to see the 1 remaining warning.

ggplot(imputed_europe_tot) +
  geom_line(aes(x = year, y = deport_convict, color = "Convict Deportation Rates"), size = 1) +
  geom_line(aes(x = year, y = deport_nonconvict, color = "Non-Convict Deportation Rates"), size = 1) +
  labs(title = "Deportation Rates in Latin America & Caribbean",
       x = "Year",
       y = "Rates per 100,000 population",
       color = "Legend") +
  scale_color_manual(values = c("Convict Deportation Rates" = "blue",
                                "Non-Convict Deportation Rates" = "green")) +
  theme_minimal() +
  theme(legend.title = element_blank())

# Filter the dataset to include only relevant variables and regions
region_data <- data_replic %>%
  filter(region %in% c("Latin America & Caribbean", "Europe & Central Asia")) %>%
  select(region, corruption, homrates_unodc)

# Visualization: Scatter plot of corruption scores vs. homicide rates by region
scatter_plot <- ggplot(region_data, aes(x = corruption, y = homrates_unodc, color = region)) +
  geom_point() +
  labs(title = "Corruption Scores vs. Homicide Rates",
       x = "Corruption Score",
       y = "Homicide Rate",
       color = "Region") +
  theme_minimal()

# Display the scatter plot
print(scatter_plot)

## Warning: Removed 370 rows containing missing values or values outside the scale range
## (`geom_point()`).

ggplot(newdf, aes(x = corruption, y = deport_convict)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region, scale = "free") +
  labs(title = "Correlation Between Corruption Scores and Convict Deportation Rates by Region",
       x = "Corruption Score",
       y = "Convict Deportation Rate")

## `geom_smooth()` using formula = 'y ~ x'

## Warning: Removed 76 rows containing non-finite outside the scale range
## (`stat_smooth()`).

## Warning: Removed 76 rows containing missing values or values outside the scale range
## (`geom_point()`).

ggplot(newdf, aes(x = corruption, y = deport_convict)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region) +
  labs(title = "Correlation Between Corruption Scores and Convict Deportation Rates by Region",
       x = "Corruption Score",
       y = "Convict Deportation Rate")

## `geom_smooth()` using formula = 'y ~ x'

## Warning: Removed 76 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Removed 76 rows containing missing values or values outside the scale range
## (`geom_point()`).

ggplot(newdf, aes(x = corruption, y = deport_nonconvict)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region, scale = "free") +
  labs(title = "Corruption Scores & Non-Convict Deportation Rates by Region",
       x = "Corruption Score",
       y = "Non-Convict Deportation Rate")

## `geom_smooth()` using formula = 'y ~ x'

## Warning: Removed 76 rows containing non-finite outside the scale range
## (`stat_smooth()`).

## Warning: Removed 76 rows containing missing values or values outside the scale range
## (`geom_point()`).

ggplot(newdf, aes(x = corruption, y = deport_nonconvict)) +
  geom_point(aes(color = region), alpha = 0.5) +
  geom_smooth(method = "lm") +
  facet_wrap(~ region) +
  labs(title = "Corruption Scores & Non-Convict Deportation Rates by Region",
       x = "Corruption Score",
       y = "Non-Convict Deportation Rate")

## `geom_smooth()` using formula = 'y ~ x'

## Warning: Removed 76 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Removed 76 rows containing missing values or values outside the scale range
## (`geom_point()`).

mean_rates <- newdf %>%
  group_by(region) %>%
  summarize(mean_homicide_rate = mean(homrates_unodc, na.rm = TRUE),
            mean_deport_convict = mean(deport_convict, na.rm = TRUE),
            mean_deport_nonconvict = mean(deport_nonconvict, na.rm = TRUE))

ggplot(mean_rates, aes(x = region)) +
  geom_bar(aes(y = mean_homicide_rate), stat = "identity", position = "dodge", fill = "blue") +
  labs(title = "Mean Homicide Rates by Region (2004-2014)",
       x = "Region",
       y = "Rate") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

ggplot(mean_rates, aes(x = region)) +
  geom_bar(aes(y = mean_deport_convict), stat = "identity", position = "dodge", fill = "red") +
  labs(title = "Mean Convict Deportation Rates by Region (2004-2014)",
       x = "Region",
       y = "Rate") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

ggplot(mean_rates, aes(x = region)) +
  geom_bar(aes(y = mean_deport_nonconvict), stat = "identity", position = "dodge", fill = "green") +
  labs(title = "Mean Non-Convict Rates by Region (2004-2014)",
       x = "Region",
       y = "Rate") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))